ILYA
@ilyaresh.bsky.social
Observability, Data, AI and tinkering with everything in between and around... https://ISbyR.com
I made #ClaudeCode & #Codex tell me when they finish a task. 🔊 #OTEL events → FastAPI → Gemini → Cartesia = spoken summaries when an agent completes. Probably abusing OpenTelemetry. Definitely useful. 😄 Blog: isbyr.com/how-i-abused... GitHub: github.com/ilyaresh/ote...
I’ve been building a #VoiceAI receptionist and discovered a whole alphabet of acronyms: DTMF, PSTN, MCP, LLM, RAG, plus the “glue” terms that connect them all. So I made a cheat sheet: The ABCs of Voice AI Agents 🎙️🤖 isbyr.com/the-abcs-of-...
The ABCs of Voice AI Agents
I present to you the ABCs of Voice AI Agents, these are some of the terms you'll stumble upon while building the Voice AI Agents
isbyr.com
I Instrumented #ClaudeCode & #Codex with #OTEL... I̴ ̴n̴e̴e̴d̴e̴d̴ ̴a̴ ̴t̴o̴p̴i̴c̴ ̴f̴o̴r̴ ̴m̴y̴ ̴n̴e̴x̴t̴ ̴L̴i̴n̴k̴e̴d̴I̴n̴ ̴p̴o̴s̴t̴.̴ why not? 😄 Sessions, token usage, costs, latency - all in #Splunk. 🔗 isbyr.com/send-claude-... #OpenTelemetry #AI
There were the #Agile and the #DevOps Manifesto(s), and now we have the “Outcome Engineering Manifesto” #O16G. o16g.com #software #development
Outcome Engineering
It was never about the code.
o16g.com
The 3 V’s of #Data still shape every modern #datastrategy: 📊 Volume – massive data being generated 🧩 Variety – structured, semi-structured, unstructured 💡 Value – turning raw data into meaningful events Get these right, and data becomes an advantage, not a costly burden.
One definitely killer use case for #AI-enabled browsers: Sitting through mandatory corporate #training and clicking “Next” the second it becomes available. 😅 #Productivity #Automation #CorporateLife
#observability hands-on time! Deploying #Splunk #O11y Cloud into one of my projects & documenting the process. What to expect: ✅ Infra #Monitoring, #RUM & #APM ✅ Real troubleshooting (not the “everything works” kind 😅) ✅ Lessons learned along the way isbyr.com/splunk-o11y-...
The results are shocking 😱 indeed! It looks like using #ChatGPT is reducing your brain inflammation and is good for your skin 👍 😀
#BOTS — #Splunk 's flagship #CTF event — is turning 10! 🎉 That’s right, Boss of the #SOC is officially a preteen now. Find out more and how to register: conf.splunk.com/connect.html #CyberSecurity #BOTS10
Great post by @jcosta.dev (@datadoghq.com) on the different #OpenTelemetry Collector 🔭 distros and when to use each. 💡 9/10 orgs use otelcol-contrib — the most bloated one, and not ideal for prod. 🔗 www.datadoghq.com/blog/otel-co... #OTel #DevOps #Observability #Telemetry@jcosta.dev
Give your #LLM clients a sense of time 🕑 with this #MCP server github.com/jlumbroso/pa...
GitHub - jlumbroso/passage-of-time-mcp: 🤖🕰️ An MCP server that gives language models temporal awareness and time calculation abilities. Teaching AI the significance of the passage of time through coll...
🤖🕰️ An MCP server that gives language models temporal awareness and time calculation abilities. Teaching AI the significance of the passage of time through collaborative tool development. - jlumbro...
github.com
#Kiro works great for #AWS apps, but hit a few bumps on a #Firebase project. Switched to #GeminiCLI to fix the issues, then let Kiro finish the job. AI tag team FTW 🤖⚡️ #AI #DevTools #CloudEngineering
@awscloud.bsky.social introduces S3 Vectors. It allows you to store your vectors at a much lower cost Vs standard vector stores, if you don't mind the increased retrieval latency. aws.amazon.com/s3/features/... #AWS #Vector #S3 #AI
Amazon S3 Vectors
aws.amazon.com
You can start contributing to #Splunk TAs! Info on how to get started is in this comment on the Splunk Ideas website. ideas.splunk.com/ideas/APPSID...
Yes, #Splunk MLTK can connect to LLMs, but the list of supported models is somewhat limited. But, you can (unofficially ⚠️) extend that model list to add support for custom models like #Google #Gemini 2.5. Details here: isbyr.com/add-new-llm-... #MLTK #LLM #AI #Observability #O11y
interesting read on how #Wiz fine-tuned a small language model to detect secrets www.wiz.io/blog/small-l... #cybersecurity #SAST #CodeAnalysis #finetuning #detection
Fine-Tuning a Small Language Model for Secrets Detection | Wiz Blog
Wiz fine-tuned a 1B LLM to detect secrets in code with 86% precision—outperforming regex-based methods while staying lean, private, and CPU-efficient.
wiz.io
Can’t believe I missed this 😲 - #Splunk MLTK 5.6.0 now supports LLMs! The new `| ai` command sends your data to a local or cloud #LLM for contextual insights 🧠 I used #DeepSeek to explain common firewall ports — got solid results 🔍 #MLTK #AI #CyberSecurity #O11y #Observability
If #Google #Gemini CLI can be used to #vibecode… Why can't it be used to build #Splunk dashboards? 🤔 Or can it?! 👀 #ToFunToStopPlayingWithIt #AI #LLM #GeminiCLI #DeveloperTools #visualisation #Observability
PSA! Just in time for Valentine's Day Exciting News! ✨ "Poem Genie" – Your go-to for crafting heartfelt Valentine's Day poems! 🌹✍️ Express love uniquely with personalized short poems. 💖 Visit poemgenie.z13.web.core.windows.net and let the magic flow! 🧞♂️✨ Dedicated to BlueSky...
My learning journey with my first #GenAI use-case, which involved a bit of playing with the #Python Pillow library and overcoming #promptpoisoning, all on #Azure. isbyr.com/my-first-gen... #development
When using #Python with #AzureOpenAI make sure to actually use your 𝐃𝐞𝐩𝐥𝐨𝐲𝐦𝐞𝐧𝐭 𝐍𝐚𝐦𝐞 for the value of the 𝐦𝐨𝐝𝐞𝐥 key.
No, #Mirabella #Genio I don't want you to create a scene, I have others around me to do that 😀 #homeautomation #smartlights
